We often hear people say, “There are so many programs out there with the same purpose, I don’t know which one to use.” One possible reason is that there are various evaluation criteria. For example, even if the performance of a method is less than or equal to that of a conventional method, it is still considered valuable if it is much faster than the conventional one. There are many differential expression methods that are more powerful under certain conditions, such as a large (or small) number of replicates in the same group. In reality, however, there are some methods whose value is questionable. This is probably due to the fact that the comparison has been biased toward representative methods. In this paper, we explain why such methods can exist, using our experience in the field of RNA-seq differential expression analysis as an example. Supplementary materials are available online at: http://www.iu.a.u-tokyo.
ac.jp/kadota/r_seq2.html#about_book_JSLAB.
View full abstract